Effects of dietary intervention on vitamin B₁₂ status and cognitive level of 18-month-old toddlers in high-poverty areas: a cluster-randomized controlled trial

The substudy was nested within a cluster-randomized, nonmasked, controlled efficacy trial conducted from March 2009 to December 2011 in a poor rural area, Xichou County, located in the Yunnan Province of China. Details of the study design have been described elsewhere [8, 9].

In the study areas, early complementary foods offered to infants are mainly various types of plant-based gruels made from cereal grains or starchy roots and tubers. Sixty administrative villages (clusters) in 9 districts in Xichou County were included in this study. Six-month-old infants were randomized to receive 50 g/d of pork (meat group), an equi-caloric micronutrient-fortified rice cereal-based supplement (fortified cereal group), or a local nonfortified rice cereal supplement (local cereal group) for one year [8, 9]. Approximately 20–30 infants were involved in each administrative village. Infants were identified at 3–5 months of age by their community doctors. Eligibility criteria for infants included term delivery without serious neonatal complications, absence of acute or chronic illness, healthy singleton status with birth weight > 2000 g, no metabolic or physical problems and being exclusively breastfed [8, 9].

Administering the intervention to participants and evaluating the compliance and morbidity of participants were performed by the specially trained community doctors in each of the small communities within each village. The fresh certified-safe lean pork was purchased weekly and was minced and accurately weighed into daily 50 g (80 kcal) aliquots containing 0.21 μg vitamin B₁₂ and stored frozen until transported weekly to the district hospitals which were serving the meat group villages [8, 9]. The quantity of either micronutrient-fortified cereal (20 g, 80 kcal) or local cereal (20 g, 80 kcal) was designed to be equi-caloric to the daily supply of pork. The participants in the fortified cereal group received a commercial product (Nestle, fortified with vitamin B₁₂, iron and zinc) and the vitamin B₁₂ content in 20 g of the fortified cereal was 0.2 μg. Local rice cereal was made from a mixture of glutinous rice flour, white granulated sugar and honey, without vitamin B₁₂ [8, 9]. Supplies of these control foods were provided weekly to the district hospitals in all participating villages, and the control foods were collected at these hospitals and distributed weekly by the participating community doctors [8, 9]. Participants in the meat group, fortified cereal group and local cereal group were encouraged to consume 50 g of red meat, 20 g of fortified rice cereal and 20 g of local rice cereal each day, respectively.

Furthermore, the participants were seen by the assessment team at the nearest hospital at baseline and at 3-month follow-up intervals. All anthropometric measurements, including length, weight, and head circumference at the age of 6 and 18 months, were measured using standardized procedures by the assessment team. The effect of supplementary food on linear growth and micronutrient status between 6 and 18 months of age was assessed. The operation manual for the study is shown in more detail in Additional file 1.

A total of 1465 infants completed the study at baseline (6 months), including 511 in the meat group, 465 in the fortified cereal group and 489 in the local cereal group. At 18 months, a total of 1316 children completed the study, including 461 in the meat group, 419 in the fortified cereal group and 436 in the local cereal group. Three children in the meat group and 15 children in the local cereal group did not provide blood samples. Finally, a total of 1298 children, without upper respiratory tract infection, diarrhea or other acute infections, provided blood samples. The flow of the participants through the study is shown in Fig. 1.

Hemoglobin (Hb), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH) and mean corpuscular hemoglobin concentration (MCHC) were measured in whole blood with an HC3000 auto-hematology analyzer on the day of blood sampling. Peripheral venous blood samples of 5 mL were collected by trained and experienced phlebotomists and put into a trace element-free, additive-free evacuated tube. Serum was divided into two aliquots and stored immediately at − 20 °C; then, they were transported frozen and stored at − 80 °C until serum vitamin B₁₂ and tHcy analysis. The individuals who carried out the laboratory analyses were unaware of the participants’ group assignments.

Serum vitamin B₁₂ was measured by an automatic ACCESS microparticle chemiluminescence immunoassay analyzer with a chemiluminescence immune assay (Beckman Coulter, Inc., USA). Serum tHcy was measured on a Hitachi 7600-120E automated biochemistry analyzer with enzymatic cycling assay (DiaSys Diagnostic Systems GmbH, Germany).

According to the vitamin B₁₂ levels of children from poor areas reported in the literature [10‐12], it is estimated that the average serum vitamin B₁₂ concentration of children in this study would be 300 pg/mL in the meat group, 350 pg/mL in the fortified cereal group, and 250 pg/mL in the local cereal group, with the assumption of a standard deviation of 150 pg/mL. Based on the sample size calculation formula for a multisample study design [13], blood samples from 60 children in each group were required to determine significant differences among the three groups and to achieve a power of 90% at a 5% level of significance. A total of 180 blood samples were collected. Sixty was randomly selected from the meat group (cluster 1–20), 60 from the fortified cereal group (cluster 21–40) and 60 from the local cereal group (cluster 41–60) by two investigators who were not involved in the recruitment and data collection, with 3 blood samples from each cluster.

The cognitive scale and the fine motor and gross motor subtests of the Bayley Scales of Infant Development, 3rd Edition (BSID III) screening test were used to assess the infants’ development at 18 months of age [14]. The language domain was not performed because it was not applicable in the setting of Yunan, China. All children were assessed within a two-week window at approximately 18 months of age, and the assessments were administered either by a psychologist familiar with the test or by trained assistants/students. All assistants/students were graduate students majoring in pediatrics who received a two-month intensive training on the BSID III test in the Department of Children and Adolescents Health Care of Xin Hua Hospital affiliated to Shanghai Jiao Tong University School of Medicine. Before the children were evaluated with the BSID III test, the evaluators had been living and working locally in rural Yunnan for several months. All evaluators were able to communicate with the participants and their parents in Mandarin and simple local dialects. Community sites were adapted and arranged to provide a standardized environment for the testing. When tests were administered, the mothers were present. If the child did not cooperate during the test, the evaluator would try again when the child was in a better mood. Each test administration took approximately 30 min. Double data entry was adopted, and the original paper copies of the data were retained for reverification.

Xichou County in Yunnan Province has an average elevation of 1400 m above sea level, so anemia was defined as Hb < 115 g/L [9, 15].

According to the World Health Organization (WHO) [12, 16, 17], serum vitamin B₁₂ concentrations of < 200 pg/mL, 200–300 pg/mL and > 300 pg/mL were used to classify individuals as deficient, marginally deficient, or adequate, respectively. However, stand-alone marker of serum vitamin B₁₂ has been proven insufficient for the unequivocal diagnosis of vitamin B₁₂ deficiency [18]. Therefore, a tHcy concentration of > 12 μmol/L was also used to obtain a correct diagnosis, based on the suggestions in the recent literature [1].

The weight-for-age z-score (WAZ), LAZ, weight-for-length z-score (WLZ) and head circumference-for-age z-score (HcAZ) were calculated according to the 2006 WHO Child Growth Standards using WHO Anthro 2011 software [19]. Underweight, stunting, and wasting were defined as WAZ < -2, LAZ < -2, and WLZ < -2, respectively.

All data were analyzed using descriptive and frequency statistics in SPSS 13.0 for Windows. The results are presented as the mean ± SD for normally distributed continuous variables or as the median (interquartile range, IQR) for data not normally distributed. Log transformations were used where residuals were skewed or exhibited nonconstant variance. Anthropometric variables, Hb level and subtest scores of the BSID III screening test were compared among groups by ANOVA. Not normally distributed data, including serum vitamin B₁₂ and tHcy concentrations and maternal education duration, were compared after log transformations among groups by ANOVA. The LSD multiple comparison test was used for the post hoc test if the overall ANOVA had a P value less than 0.05. Chi-square tests were used for categorical variables. To address the clustering effect, the differences in serum vitamin B₁₂ and tHcy concentrations (after log transformations) and cognitive score between the meat or fortified cereal groups and the local cereal group were analyzed using mixed-effect linear models, and the local cereal group was used as the reference. Outcomes were analyzed with villages as the random factor and group as the fixed factor. Comparisons were expressed as estimated differences with 95%CIs. Correlations between serum vitamin B₁₂ and tHcy levels and the associations of vitamin B₁₂ or tHcy levels with LAZ, WAZ, WLZ, HcAZ, Hb, MCV, MCH and MCHC were tested by Pearson correlation after log transformation. The associations between the subtest scores of the BSID III screening test (the cognitive score, the fine motor score or the gross motor score) and the serum vitamin B₁₂ or tHcy concentrations were examined by a linear regression model on condition of the adjusting covariates (maternal education duration and birth weight) with “enter” as the regression method. All statistical tests were two-tailed and P values < 0.05 were considered statistically significant.

A total of 180 blood samples, 60 from each group, were selected by a simple random sampling from 1298 blood samples stored at − 80 °C in Shanghai Key Laboratory of Children’s Environmental Health.

The mean age of toddlers in the meat group, fortified cereal group and local cereal group was 17.9 ± 0.1, 18.0 ± 0.2 and 17.9 ± 0.2 months, respectively. The percentages of males among the three groups were 51.7, 48.3 and 48.3%, respectively. The mean birth weights in the three groups were 3075.7 ± 380.2, 3039.6 ± 362.1 and 3005.7 ± 373.4 g, respectively, and the median maternal education duration in the three groups was 9(6–9), 9(6–9) and 9(6–9) years, respectively. There was no significant difference in birth weight and maternal education duration; furthermore, the LAZ, WAZ, WLZ and HcAZ at 18 months of age were similar among the three groups (all P > 0.05) (Table 1).

The median (IQR) concentration of serum vitamin B₁₂ was 360.0 (233.0–573.8) pg/mL in all children at 18 months of age. Deficient (< 200 pg/mL) and marginally deficient (200–300 pg/mL) serum vitamin B₁₂ concentrations were found in 62 (34.4%) children and 30 (16.7%) children were vitamin B₁₂ deficient.

The median (IQR) level of serum tHcy was 8.2 (6.9–10.2) μmol/L in all children. Serum tHcy > 12 μmol/L was found in 21 (11.7%) children, and among them, 7 children also had serum vitamin B₁₂ levels < 200 pg/mL, and another 7 children had serum vitamin B₁₂ levels of 200–300 pg/mL.

The mean Hb level in all children was 122.5 ± 11.8 g/L. Anemia (Hb < 115 g/L) was found in 42 (23.3%) children, and among them, 14 children also had serum vitamin B₁₂ levels < 300 pg/mL. There was no significant difference in Hb, MCV, MCH and MCHC among the meat group, fortified cereal group and local cereal group (P > 0.05) (Table 1).

Pearson correlation results suggested that serum vitamin B₁₂ levels were negatively correlated with tHcy after log transformation (r = − 0.45, P < 0.001). There was no association between vitamin B₁₂ level and the LAZ, WAZ, WLZ, HcAZ, Hb, MCV, MCH or MCHC (P > 0.05). The tHcy level was positively correlated with the WLZ after log transformation (r = 0.15, P = 0.039), but there was no correlation between the tHcy level and the LAZ, WAZ, HcAZ, Hb, MCV, MCH or MCHC (P > 0.05).

The median vitamin B₁₂ concentration was significantly different among the three groups (P < 0.001). The fortified cereal group had the highest level (509.5 pg/ml), followed by the meat group (338.0 pg/ml) and the local cereal group (241.0 pg/mL) (Table 1). Compared with the local cereal group, toddlers in the meat group (estimated difference 0.16, 95% CI: 0.06~0.26, P = 0.002) and fortified cereal group (estimated difference 0.35, 95% CI: 0.26~0.45, P < 0.001) had higher vitamin B₁₂ concentrations, according to the mixed-effect linear models after log transformations.

Among 62 children whose vitamin B₁₂ concentration was < 300 pg/mL, there were 22 (36.7%) in the meat group, 3 (5.0%) in the fortified cereal group and 37 (61.7%) in the local cereal group. The difference between the three groups was statistically significant (χ² = 42.86, P < 0.001).

There was a significant difference in the serum tHcy levels among the three groups (P < 0.001). The serum tHcy levels in both the meat group and fortified cereal group were lower than the levels in the local cereal group (P < 0.05) (Table 1). Compared with the local cereal group, toddlers in the meat group (estimated difference − 0.08, 95% CI: − 0.13~ − 0.02, P = 0.005) and fortified cereal group (estimated difference − 0.12, 95% CI: − 0.18~ − 0.07, P < 0.001) had lower tHcy concentrations, according to the mixed-effect linear models after log transformations.

Among 21 children whose tHcy level > 12 μmol/L, 7 (11.7%) were in the meat group, 3 (5.0%) were in the fortified cereal group and 11 (18.3%) were in the local cereal group. There was no difference among the three groups (χ² = 5.18, P > 0.05). Serum tHcy > 12 μmol/L and vitamin B₁₂ < 300 pg/mL were found in 14 (7.8%) toddlers. Among them, 7 toddlers had a vitamin B₁₂ level of < 200 pg/mL.

Because 9 children were unable to cooperate with staff to complete the test, including 2 in the meat group, 5 in the fortified cereal group and 2 the in local cereal group, 171 children completed the BSID III screening test at 18 months of age.

There were significant differences in the cognitive score of the BSID III screening test among the three groups (P = 0.020). Compared to the local cereal group, children in both the meat group and fortified cereal group had higher cognitive scores (P < 0.05) (Table 2). The mixed-effect linear models showed that the meat group had higher cognitive scores than the local cereal group (estimated difference 0.99, 95% CI: 0.21~1.76, P = 0.013). The cognitive score of the fortified cereal group was also higher than that of the local cereal group, but the difference was not statistically significant (estimated difference 0.78, 95% CI: − 0.00~1.57, P = 0.051).

In 171 children, serum vitamin B₁₂ concentrations were positively correlated with the cognitive score (beta = 2.15, SE = 0.55, P < 0.001) and the fine motor score (beta = 0.71, SE = 0.31, P = 0.023), according to a multiple linear regression analysis.